Slashdot Mirror

Linux makes a great desktop for people who have a high level of computer "know-how" or programmers but lacks the consistency and polish need to be anyone's desktop.

What GNU/Linux needs in order to be truly desktop ready is consistency between distributions, specifially (in no particular order):

1. A common pacakage format. RPM and DEB are both nice but the Linux Standards Base (LSB) has decided on RPM so let's all just use it. RPM is open source software, if it doesn't have something that DEB does, suck it up and add it.
2. A consistent configuration layout. I should be able to log onto Gentoo, Fedora, RHEL, Debian or Ubuntu and find that the configuration files for a given application or system feature are the _same_ format and exist in the _same_ places on the system. Why does every distro need to do something this simple in 10 different ways?
3. A solid set of GUI configuration tools. Following the previous point about a consistent configuration layout, there should be a common API for manipulation of said files and a set of GUI applications that can do common tasks without requiring a user to drop to the command prompt.
4. A package manager capable of installing software that the user downloads from the internet. Users of Mac OS and Windows expect to be able to go to Evolution's - for example - website and download and install the latest version of that application. This _has_ to work.

Linux needs a universal framework for application development

app installation/uninstallation

user configuration storage

sound and graphics

users won't have to install two entire desktop environments just to run all the apps out there

I prefer windows because I get notified in a uniform manner when hardware is added. I prefer windows because there aren't fifty ways to access the same device. I prefer windows because it's easier to develop software for. Have you ever written dynamic hardware detection software for Linux/Solaris?

Don't need to. We have HAL/D-Bus now.

If you have to recompile the god damn kernel just to see your storage it's not user friendly.

Most distros provide kernels pre-compiled with everything available as modules. D-Bus auto-loads modules as needed when new hardware is detected.

Windows is like a Honda Accord, not very exciting, but it works, Linux is like a Ferrari...that's mailed to you in pieces over the course of a year, each piece comes with a manual of how it works and what it does, but gives no instruction on how to use it with other components.

There're these new-fangled things called "desktop distributions." They're mail-order Ferraris that cost less than a Honda Accord (most even free) that come pre-assembled with a dashboard computer that asks you your name when you first get in and helps you tailor itself to your needs. You can even fix the engine yourself if you're so inclined - unlike that Honda that comes with non-standard, esoteric parts that amaze you that it works every time you think you've figured out how it's put together; and you'll be threatened with a lawsuit or jail time if you try to change anything under the hood.

I prefer windows because I get notified in a uniform manner when hardware is added. I prefer windows because there aren't fifty ways to access the same device. I prefer windows because it's easier to develop software for. Have you ever written dynamic hardware detection software for Linux/Solaris?

Don't need to. We have HAL/D-Bus now.

If you have to recompile the god damn kernel just to see your storage it's not user friendly.

Most distros provide kernels pre-compiled with everything available as modules. D-Bus auto-loads modules as needed when new hardware is detected.

Windows is like a Honda Accord, not very exciting, but it works, Linux is like a Ferrari...that's mailed to you in pieces over the course of a year, each piece comes with a manual of how it works and what it does, but gives no instruction on how to use it with other components.

There're these new-fangled things called "desktop distributions." They're mail-order Ferraris that cost less than a Honda Accord (most even free) that come pre-assembled with a dashboard computer that asks you your name when you first get in and helps you tailor itself to your needs. You can even fix the engine yourself if you're so inclined - unlike that Honda that comes with non-standard, esoteric parts that amaze you that it works every time you think you've figured out how it's put together; and you'll be threatened with a lawsuit or jail time if you try to change anything under the hood.

http://freedesktop.org/wiki/Xming

Might I suggest that we work on creating good Nvidia Linux drivers before we work on overclocking our poorly driven hardware. I'm counting on nouveau http://nouveau.freedesktop.org/wiki/ to save us unix/nvidia users in this respect.

1. GTK 2.8 was no great jump over GTK 2.6 -- unlike Vista over win 3.1.

So you find it acceptable in some cases then.

2. There's nothing stopping you from applying the patches that fix bugs yourself. - the war cry of the moron whenever FOSS fucks up really badly.

If your shipping a few million PCs then I'm sure it's not going to be too much overhead for you.

3. Cairo should use hardware acceleration so it shouldn't be so much of a problem if you had a decient PC. -- Cairo doesn't use hardware acceleration.

Yes it does it's called glitz

4. The OLPC isn't a decent PC... fuckwit

see 2

5. Wasn't one of the selling points of FOSS the fact that older machines don't get left behind and put on forced upgrade treadmills.

see 2

You're a fucking idiot, mate.

and your a ludite.

And, there are a very restricted number of video cards with good open/free drivers for them. (ATi cards with up to and including r300 have pretty good drivers (GL 1.0 is supported with the Mesa driver) and there are reports that the nouveau drivers included in FC6 are actually working now for some nVidia cards.

But in general, gaming sucks on Linux because the video card manufacturers have made it hard to get accelerated graphics working with the open standard, then on top of this is your very valid point about DirectX.

Yes, they could have, but no they haven't as this would violate Nvidia's license and you wouldn't be able to distribute the driver once you put in all that work. But see http://nouveau.freedesktop.org/wiki/ for a project to legally reverse a driver for Nvidia cards.

Additional comment #2 from https://bugs.freedesktop.org/show_bug.cgi?id=3654:

"When using the proprietary nvidia driver the same problem happens, only that the
error message (Mode "1920x1200" is larger than BIOS programmed panel size of
1024 x 768. Removing.) does not appear."

Btw, I'm running 1920x1200 through a DVI cable with the nv driver, and have done so for at least a year...

I'm not exactly sure what you are asking, it doesn't even really sound like you are sure what you're asking. Anyway take a look at: http://freedesktop.org/wiki/Standards which is pretty much standards for running KDE apps on Gnome and vice versa (as well as some other issues). At this point Gnome and KDE apps cooperate reasonably well with one another and KDE works as well Gnome as OO does so ....

Well, then they could just release the parts that they have rights to and let the open source hackers fill in the remainder. It doesn't seem like that would be a particular challenge. Also these guys are working on open source 3D drivers for nVidia cards. We should probably give them some more help.

The nouveau project is actively working on a free software driver for nVidia cards that will hopefully replace the nv driver one of these days. They could use some help.

http://nouveau.freedesktop.org/wiki/
http://wiki.x.org/wiki/nv

Open Source 3D capable drivers for nvidia are under development by the Nouveau Project , stay tuned or just help them out!

"Score: 4 Informative" or just all the people knows that?

That is because no one used nvnet.

The SAME thing has basically happened with r200 and fglrx. The r200 driver surpassed fglrx in quality and performance so it made no sense to continue to support r200 in fglrx!

Trade secrets seem like a bad argument if you can't get anyone to use your supposed "programmatically advanced" code.

Everyone should just support nouveau:
http://nouveau.freedesktop.org/wiki

* Intel integrated graphics. (All of them. Intel actively provides Free Software drivers for their graphics chipsets.)
* ATI Radeon <r500 (that's <= X850, including all the non-X ones), with the exception of the shared RAM Xpress integrated chipsets (If you want an integrated chipset, pick Intel!). http://dri.freedesktop.org/wiki/ATIRadeon

https://bugs.freedesktop.org/show_bug.cgi?id=3654

"The "nv" driver currently can't change the BIOS-programmed display timings. Unfortunately, this is not something that we can fix right now."

This just sucks, IMHO.

I've been AMD/NV since the athlon launch and I'm looking at switching to intel. This clinches the deal.

AMD has dropped the ball and ATI never even got close to the ball.

it'll also put the OK and Cancel buttons the right way around.

Well,it'll allow toolkits to put them the right way around for the desktop you're on - version 1.6 of the API doc has a ButtonOrder() API to let a toolkit determine the appropriate button order for the desktop on which it's running.

If that's not what you consider the right way around, either switch to a desktop that puts them in the order you want, try to get your preferred desktop to put them in the order you want, or try to get them to offer an option to control the button order. Portland doesn't exist to standardize the look and feel of desktops, it exists to allow applications written for one desktop to work better on other desktops.

it'll also put the OK and Cancel buttons the right way around.

Well,it'll allow toolkits to put them the right way around for the desktop you're on - version 1.6 of the API doc has a ButtonOrder() API to let a toolkit determine the appropriate button order for the desktop on which it's running.

If that's not what you consider the right way around, either switch to a desktop that puts them in the order you want, try to get your preferred desktop to put them in the order you want, or try to get them to offer an option to control the button order. Portland doesn't exist to standardize the look and feel of desktops, it exists to allow applications written for one desktop to work better on other desktops.

OK, You mean, like, CDE? :)

No, not unless CDE adopts it, which is probably pretty unlikely at this point (I'm not sure anybody's actively developing it).

However, the press release nonwithstanding, this is not intended solely to be used by KDE and GNOME; the FAQ lists XFCE as another probable supported desktop environment.

"If it's not already in your development tree or toolkit, xdg-utils is available for download at http://portland.freedesktop.org/wiki/. "

The great thing about standards is there are so many to choose from...

As long as the webpage is sufficiently standalone encapsulated, I can drag its URL into my Ubuntu Panel, where it makes a button I can click like any other app.

If I wanted, I could write an HTML wrapper I keep on my local machine or my own webserver that pops up Javascript UIs to populate the URL with parameters for opening the remote webpage.

The only real problem is IPC between the webpage app, but that's always been a terrible problem with webpages since the beginning that practically no one has addressed. Maybe as these remote apps are opened in suites with each other (and with local apps) the demand will force a better IPC, probably according to some FreeDesktop.org standard.

Your turn.

(avi)-- http://people.freedesktop.org/~davidr/xgl-demo1.xv id.avi

And Compiz can run on older equipment as well.. :)

http://en.wikipedia.org/wiki/Compiz

-- NOOOO!! I am not trying to get into the 'good graces' of /. management!! :]

Actually Linux is now equal to and seemingly pulling ahead of Macs, especially in the graphics areas..

http://en.wikipedia.org/wiki/Xgl

And some real magic..

(avi)-- http://people.freedesktop.org/~davidr/xgl-demo1.xv id.avi

-- NOOOO!! I am not trying to get into the 'good graces' of /. management!! :]

http://lists.freedesktop.org/archives/xdg/2006-Apr il/008012.html
http://lists.freedesktop.org/archives/xdg/2006-Apr il/008025.html
http://lwn.net/Articles/178411/
http://lwn.net/Articles/178409/

http://lists.freedesktop.org/archives/xdg/2006-Apr il/008012.html
http://lists.freedesktop.org/archives/xdg/2006-Apr il/008025.html
http://lwn.net/Articles/178411/
http://lwn.net/Articles/178409/

KDE and GNOME merging would be a horrible idea. First, for technical reasons: GNOME is constructed on GTK, mostly in C (and, lately, C# / Mono). KDE is built on QT and mostly C++. If both projects somehow merged, you'd either end up with having to throw one of the toolkits out of the window, or you'd end up having a desktop environment that requires two toolkits. Yes, you might argue that most modern Linux desktops have both toolkits installed already anyway. But a desktop environment that is built on more than one base just wouldn't make a lot of sense and would most likely be cumbersome to develop for. The fact that you'd have two programming languages (albeit related ones) probably wouldn't help that much, either.* And as for just dropping one of the toolkits and languages - you'd have to more or less re-write roughly a gazillion applications from scratch just to fit into the new scheme. Hmm. I've heard of better ideas. You could say the same about Human Interface Guidelines. You'd definitely need to throw one set out of the window, and make non-trivial changes to the interfaces of countless programmes. * Yes, I do know that other languages are occasionally involved, and some projects already do mix C and C++ in some ways. I'm simplifying things a bit here on purpose. Then, there's the fact that the two projects follow a completely different design mentality. Now, me, personally, I've never liked GNOME much. There's some elements I like and it's definitely "tidier" than KDE in many ways and there's lots of delicious eyecandy, but there's also tons of stuff I dislike. I hate that stupid menu bar at the top, I hate the awful file dialogues, I hate the way preferences are dumbed down, I consider Nautilus to be useless both in spatial and in browser mode, gEdit is a sorry excuse for a text editor in my eyes. KDE on the other hand - I love the way I can set it up just the way I like it, I find it much more usable than GNOME, IOSlaves and the high degree of integration between the different programmes have made my life much easier, JuK is a nice little jukebox, Konqueror a gorgeous file manager and web browser, Kopete is much nicer than gAIM nowadays, Kate is the best text editor I've ever seen - and so on, and so on. Remember, that's just my personal preference, I'm definitely not trying to start a flamewar. Because for everyone who agrees with me, there's someone who prefers GNOME for his own personal reasons and hates KDE. And that's OK. Choice is one of the best things about FOSS. Somehow merging GNOME and KDE would destroy much of that choice. So, do I advocate the Linux desktop being fragmented forevermore? Actually, I don't, even though I think a KDE/GNOME merger would be terrible. I think what we need is not one unified desktop, we need standards that most Linux desktops adhere to. We need the different desktops to be as interoperable as possible. To a large degree, they already are - I can run gAIM or GIMP under KDE just fine. And things are getting better still. If you haven't heard of it yet, you should check out the impressive desktop standardisation work of The Portland Project. To sum it all up: A merger would be near-impossible to pull off and a crappy idea because it requires lots of re-writing, destroys choice and makes things more difficult for developers. However, common standards are extremely useful. IMHO, the Linux desktop is already heading the right way as it is, no matter whether you use GNOME or KDE or something else!

Oh I believe that you are correct regarding the shadows. But I'm not a graphic designer and I never noticed something like that. And I'm willing to bet that most people here on Slashdot are the same. So while your comments are valid I don't think anyone here cares. To find someone that cares you'd need to go to, you guessed it, forums for projects like Tango.

And when coders comment on code (which I rarely see here on Slashdot) or someone comments on inaccuracies in Wikipedia or someone finds a bug in a program what do everyone here inevitably say? "Did you fix it or file a bug report?"

Personally I find people who complain and then don't do anything about it just plain arrogant. Nobody is expecting that you give a free seminar to them. But perhaps you should consider that many working on OSS projects actually have high-paying jobs. Personally I find it rewarding enough to share what I know with other people. And since I don't know everything (yet) it gives me an opportunity to learn new stuff from other people at the same time.

BTW if you had contacted them or looked at their mailing-list you would have found that they already know. The reason they are not doing it right now (and that there is no shadow policy in their guidelines) is that the OSS toolkit doesn't support a good way of making them.

For the meantime we avoided talking about shadows as the free software
tools lack in that area and we are still using workarounds, rather than
proper solutions to provide object shadows. Once the free software stack
catches up (Inkscape mainly, coming real soon), I'll add a section on
shadows.

http://lists.freedesktop.org/archives/tango-artist s/2006-September/000664.html

Tango might want to work a bit on there web site security! (it's been hacked, apparently by somebodying editing all the lead wiki pages to say "HA HA").

I'll take part of the bait - anonymously anyway...

Compiz - nothing else comes close to virtual desktop in 3D, Video demonstration of Compiz on Xgl (linked at the bottom of wikipedia page direct to video)

I've been using Windows since 95 on a P1 all through 98, NT4, ME, 2000, XP, and my current XP2003 with 200 and 300 gig hdd's, Athlon 64 X2 4200, XFX GeForce 7900 GT, 4 gigs of ram, and twin/dual display 17" lcd's, - and Linux is great...

BTW: did I mention I also put Linux on my laptop (an Athlon 3200 with a gig of ram). Not everybody that runs linux cheaps out on hardware - in fact, more windows users cheap out on machines than linux - what can you say about Dell's running XP? not shit. On the other side, what do you have that can compete with supercomputers such as the one this article is supposed to be about - IBM's next 1.6 petaflop supercomputer - which will run what? LINUX!!!

That's not really true, considering the nvidia open source drivers do not even support 3d acceleration. The ati open source drivers are alot further along but only support older cards.

If you have a newer card or a nvidia card, the only option is XGL/compiz which has the same effects (and more) than the new Metacity. If you still want to use Metacity you will have to wait until Nvidia/ATI releases their drivers with texture from pixmap support which could be 6 months to a year from now. XGL has tfp already built into its server which allows one to use accelerated 3d effects even if their driver does not support it.

Also I should note that one could use compiz with AIGLX (not sure if you can right out of the box or requires a patch).

I remember years ago when Gnome was the eye-candy window manager all the kids were showing off. In looking through the screenshots, the most surprising thing is to see that nobody involved with the Tango interface has ever seen what an actual shadow looks like.

If you want to do flat shadows, cool, do them, they're easy and effective. If you want to do three-dimensional shadows, cool, they look even better but take a bit more work. But don't drop the same blurry ellipse at the bottom of every object and think that you're making a three-dimensional shadow, you just make everything look like it's standing on a blurry gray oval, and users really do recognize the less professional look, consciously or not.

Compiz is still early in development (buggy and hard to install) and undergoing constant changes, but it finally makes Linux look competitive on the desktop. Without Compiz, Linux is the blandest of 3, even though you can easily modify themes, it doesn't help much. With it, it looks better than OS X. It even works on 5 year old video cards (unlike Vista). I hope in 6-12 months, Compiz will be stable enough to be default in major distributions.

Linkz:
http://en.wikipedia.org/wiki/Compiz
http://www.freedesktop.org/~davidr/xgl-demo1.xvid. avi
http://www.youtube.com/v/DUSn-jBA3CE
http://compiz.blogspot.com/

Why are they fleeing gstreamer like plague? I see at least three reasons:

1. Because gstreamer's performance sucks; there is absolutely no excuse for using ten times more resources than mplayer/xine for decoding a MP3;
2. Because after 7 years (yes, the project started even before Windows XP and OS X were released!), it still is brittle (look yourself at the serious bugs fixed two weeks ago);
3. Because no serious video player uses it (there must be a reason...); and even in simple audio apps, see point 1. Why would Adobe be a testbed for a library that didn't prove to be reliable video-wise?

No, that's different. Read the ATI section of the DRI website. Specs for cards up to the Radeon 9200 are available to individual developers, pending an NDA. That is *not* an open API. I don't know what technology is included in the developer resources, but if it's protected by NDA, I would have to assume some part of it are considered by ATI to be trade secrets.

Check out Mike Harris' take on the NDA/DRI driver development issue. I know, this doesn't reference ATI specifically, but the ATI section of the FDO wiki does, unequivocally, state that the important reference materials are locked up. Anyway, from Mike's treatment of the subject, it is clear that not just anybody gets to see the guts. I'm assuming that the NDA-restricted information excludes actual licensed IP (SGI's and S3's).

Yeah, the S3TC situation is crap. In that case, the text compression format is well known and documented. We're just not allowed to implement it because the right to use it would have to be licensed directly from S3 (or whomever it is that owns them now). It is a problem.

They should release it for non-commercial use, at least.

Who knows where that licensing authority is now... could be the guys that bought SONICblue, could be VIA, etc. S3 kinda exploded, and the pieces flew out pretty far. I wouldn't be surprised it ATI and nVidia picked up bits, too.

This seems like a good on-topic thread in which to mention the freedesktop.org (X.org folks) effort to write a 100% open source 3D driver for the NVidia cards -- nouveau

http://nouveau.freedesktop.org/wiki/

If you're an owner of an nVidia card, please do all you can to help contribute! They appear to be suprisingly far along.

--
Slashcode bug # 497457 - unfixed since December 2001 - Go look it up!

No, the X11 protocol does not need confirmation for every request; it's Xlib's design which makes it synchronous. The xcb library has been designed as an asynchronous replacement for it, and does not have that problem, while using the same protocol.

I'm no expert in this

Indeed. :-) Actually, the chattiness is more a fault of xlib (libX11, the C binding to X11) than the X Protocol's design. So much so that there's a project making a modern alternative to xlib called XCB http://xcb.freedesktop.org/wiki/- still uses the exact same network protocol, it's just a better designed C API to it!

This seems like a good on-topic thread in which to mention the freedesktop.org (X.org folks) effort to write a 100% open source 3D driver for the NVidia cards -- nouveau

http://nouveau.freedesktop.org/wiki/

If you're an owner of an nVidia card, please do all you can to help contribute! They appear to be suprisingly far along.

--
Slashcode bug # 497457 - unfixed since December 2001 - Go look it up!

I agree that case-insensitivity is nice when working in English; however, in Arabic, short-vowel insensitivity would be more useful. In German, you have the complexity of ß versus SS versus ss. French requires you to ignore some accents. In short, the rules for each language are different, and on Windows 98, I had a few problems with foreign colleagues creating file names I couldn't handle (until we agreed on all-uppercase, English only names), because the two filenames were identical to Windows 98 English edition, but not to their foreign edition.

The point at which case sensitivity should be fixed is the point at which an application is prompting for a filename, not lower. If you're using tab-completion, I can do (locale-specific) processing to determine other plausible filenames, and print them for you to choose from. If you're using a GUI application (far more common), I can do case-insensitive handling of typed filenames, and highlight the file you've selected in a dialog box. I can even prompt you when you're creating two files with names that only differ in case, so that you can't do that accidentally. Any lower level, and you get into trouble with different languages having different case sensitivity rules.

Pretty please do spend the time to report the small problems which keep you from using DejaVu on
https://bugs.freedesktop.org/enter_bug.cgi?product =DejaVu

As you wrote the DejaVu project will fix these small issues in time, but this time depends on informed people like graphic designers reporting what needs to be fixed exactly.

standards

CLI programs are REALLY useful to look at "hidden" content.

'pdftotext' comes with xpdf & is even available natively on windows.

Similarly, for MS Word documents, you may use 'antiword', 'catdoc', and 'wv'.

These programs are quite nice in that they can easily batch-process a lot of documents & then you can go grepping through them for interesting tidbits.

(On the GUI front, evince deserves a plug. It uses the same poppler backend as xpdf and kpdf. I used to use tiny & fast xpdf for most of my pdf viewing, but evince has a few nice features which xpdf lacks & has become my personal favorite pdf viewer.)

GNOME is steering away from CORBA, to the point of rewiring their next major release for D-BUS. I believe this includes Bonobo.

The UNIX/Linux world has never really had a good way for applications on the same machine to intercommunicate in a subroutine-like way. Microsoft has OLE/COM/DCOM/ActiveX, which is clunky but always available. In the Linux/Unix world, there's nothing you can assume is always there. There's OpenRPC, there's CORBA (in about five incompatible forms), there's Java RMI, and there are various kludges built out of pipes. But there's been no convergence in two decades.

D-BUS

the Free driver really isn't worth much.

There are no OSS nVidia drivers at all. None. Nope.
Stuck with the binary-only drivers.

ATI's binary driver may be pure shit (as far as I've heard from other users) but :
- there's some opensource DRI support.
- there're some reverse engeneering efforts done to support more recent cards.

So, monomaniac opensrouce zealot, like myself, are more likely to be happy from a merger between those to companies.

Also didn't ATI have plans to use HyperTransport for their graphic chips as opposed to nVidia only using it for chipset and implementing their own protocol for SLI ? I'm not sure, I'm only trying very hard to remember what I've read. Can anyone help ?

Well, you do have to lay some (I'd argue most) of the blame on the toolkit or application authors. For example, forget the docs, just look at the prototype for XRenderCompositeString8. Does that prototype suggest to you that you should draw an entire page full of text a single letter at a time using that function? 'cause that's what at least one text editor does.